CVE Watch399,240 in full archive

Vulnerabilities exploitable today

14,859in current view

Single score combining CVSS, KEV membership and EPSS. Every CVE with its own record — timeline from publication to active exploitation.

In KEV catalog1,748

New KEV · 24H0

Exploit Today ≥ 701,665

Distribution · last window

Critical
3,347
High
15,203
Medium
14,178
Low
1,360

Filters

Window

Severity

Flags

Vulnerabilities13,801–13,840 · 14,859

CVECVSSEPSSKEVRExploitTitleMod.

CVE-2026-844868.2 HIG

43.3%

——13A flaw was found in Red Hat Ansible Automation Platform's automation- controller. Four debug views that trigger the internal task, dependency, and workflow schedulers are configured to allow any user (including unauthenticated clients) and are routed in production builds because their URL include is not gated on the debug setting. An unauthenticated remote attacker can repeatedly invoke these endpoints to acquire the cluster-wide scheduler advisory lock; because the legitimate scheduler acquires the same lock without waiting, the attacker causes real scheduler runs to be skipped, stalling job dispatch for all tenants, while also consuming controller web workers. The debug root view additionally discloses the list of debug endpoints to unauthenticated callers.27d CVE-2026-844749.9 CRI

55.9%

——17A flaw was found in Red Hat Ansible Automation Platform's automation- controller. The provisioning-callback secret (host_config_key) is exposed to users holding only the read-level view_jobtemplate permission -- both in the job template API representation and in the activity stream -- and the provisioning callback endpoint trusts a client-supplied X-Forwarded-For header to determine the calling host when the controller is deployed behind the AAP gateway with an empty proxy allow-list. By reading the secret and spoofing X-Forwarded-For to match any host in the job template's inventory, a minimally privileged or unauthenticated remote attacker can launch the job template against arbitrary managed hosts using the job template's credentials, resulting in privilege escalation and remote code execution on managed hosts.27d CVE-2026-82368—

15.8%

——5Insecure access controls on internal service ports in Brocade SANnav versions before 3.0.1a allow local, non-administrative host users to communicate directly with backend management services. A local attacker can leverage this exposed access to transmit commands to connected Fabric OS switches under the security context of the SANnav management user.19d CVE-2026-823567.5 HIG

1.2%

——0Imprivata EAM <=26.2.6 lacks the ability to rotate its RSA key pair after deployment when generating an X.509 certificate. Using an RSA key pair indefinitely for certificate generation is against best practices.25d CVE-2026-776029.9 CRI

46.3%

——14OpenC3 COSMOS provides the functionality needed to send commands to and receive data from one or more embedded systems. From 5.1.0 until 7.3.0, authenticated non-administrator users can write content under targets_modified/ that is later executed by multiple configuration paths below the intended code-execution privilege tier. Table and command or telemetry definitions are processed through ConfigParser, PacketConfig, GENERIC_READ_CONVERSION, or GENERIC_WRITE_CONVERSION, allowing ERB rendering or Ruby and Python evaluation, while openc3-cosmos-script-runner-api/scripts/run_suite_analysis.rb executes suite procedure files through require. Storage uploads, screen saves, and script creation can place content in the overlay, and triggering table processing, a cmd/tlm reload, or suite analysis executes the content in cmd-tlm-api, decom microservices, or Script Runner with access to internal credentials and data. This issue is fixed in version 7.3.0.22d CVE-2026-776018.8 HIG

46.7%

——14OpenC3 COSMOS provides the functionality needed to send commands to and receive data from one or more embedded systems. From 5.12.0 until 7.3.0, an authenticated actor can write the pypi_url setting through set_setting at POST /openc3-api/api, then cause OpenC3::PluginModel.install_phase2 in openc3/lib/openc3/models/plugin_model.rb to interpolate the value into a shell command while installing a plugin with Python dependency metadata. Shell metacharacters in the setting are interpreted by the command shell, allowing arbitrary operating-system commands to run as the openc3 service user with access to Redis and bucket credentials. Open-source deployments permit any authenticated user to reach the affected operations, while Enterprise deployments require an administrator. This issue is fixed in version 7.3.0.27d CVE-2026-774237.5 HIG

31.0%

——9JLine is a Java library for handling console input. From 3.0.0 until 3.30.15 and 4.3.1, the JLine built-in less viewer passes user-controlled search and display-filter patterns from getPattern(boolean doDisplayPattern) in builtins/src/main/java/org/jline/builtins/Less.java directly to Java's backtracking regular expression engine and repeatedly applies them to file content. A nested-quantifier expression evaluated against non-matching lines can consume excessive CPU and indefinitely block the session thread, and repeated sessions in Telnet or SSH deployments can exhaust a bounded worker pool. This issue is fixed in versions 3.30.15 and 4.3.1.26d CVE-2026-774227.5 HIG

41.3%

——12JLine is a Java library for handling console input. From 3.0.0 until 3.30.15 and 4.3.1, the JLine built-in grep command in builtins/src/main/java/org/jline/builtins/PosixCommands.java accepts a user-controlled regular expression in grep(...) and, unless line-regexp mode is used, automatically adds a dot-star prefix and suffix before compiling it with Java's backtracking regular expression engine. The wrapping expands the backtracking search space, so a short nested-quantifier expression evaluated against non-matching input can consume excessive CPU and indefinitely block a command worker, including in remotely exposed shell sessions. This issue is fixed in versions 3.30.15 and 4.3.1.26d CVE-2026-774216.5 MED

36.3%

——11JLine is a Java library for handling console input. From 3.0.0 until 3.30.15 and 4.3.1, the JLine built-in nano editor's regex search mode passes a user-controlled search term from doSearch(String text) in builtins/src/main/java/org/jline/builtins/Nano.java to Java's backtracking regular expression engine without a timeout or backtracking bound. A nested-quantifier expression evaluated against non-matching buffer content can consume excessive CPU and indefinitely block the editor session thread, and remote multi-user deployments can lose a worker thread for each affected session. This issue is fixed in versions 3.30.15 and 4.3.1.27d CVE-2026-774205.5 MED

1.9%

——1JLine is a Java library for handling console input. From 3.0.0 until 3.30.15 and 4.3.1, DefaultHistory.matchPatterns(String patterns, String line) in reader/src/main/java/org/jline/reader/impl/history/DefaultHistory.java converts the HISTORY_IGNORE configuration value into a Java regular expression while escaping only part of its syntax, allowing other regex metacharacters to reach the backtracking engine. An attacker who can control application or user configuration can supply a nested-quantifier expression that is reevaluated whenever a command is added to history, consuming excessive CPU and indefinitely blocking the reader thread. This issue is fixed in versions 3.30.15 and 4.3.1.22d CVE-2026-773947.6 HIG

31.6%

——9OpenC3 COSMOS provides the functionality needed to send commands to and receive data from one or more embedded systems. From 5.0.6 until 7.3.0, an authenticated actor with system_set permission can store a shared screen through POST /openc3-api/screen whose BUTTON widget action is evaluated by openc3-cosmos-init/plugins/packages/openc3-vue-common/src/widgets/ButtonWidget.vue in another operator's browser session when the button is activated. The stored script runs in the COSMOS origin and can read localStorage.openc3Token, allowing theft of the victim's bearer token, account takeover, and actions with the victim's privileges. The permissive content security policy contributes to execution but is not the primary root cause. This issue is fixed in version 7.3.0.27d CVE-2026-77285—

2.0%

——1OpenBao is an open source identity-based secrets management system. Prior to 2.6.0, OpenBao Agent's exec rendering mode could write secrets from env_template to standard output when command/agent/exec/exec.go re-created the template runner after repeated rendering failures, primarily after num_retries was reached. A process supervisor, log collector, or local user able to read that output could obtain the rendered secret values. This issue is fixed in version 2.6.0.21d CVE-2026-766488.5 HIG

13.7%

——4CopyAPIView (awx/awx/api/generics.py:873) sets permission_classes = (IsAuthenticated,), so DRF's get_object() performs no object-level RBAC. The get() handler (lines 988–991) explicitly guards with request.user.can_access(obj._class_, 'read', obj) — but post() (lines 1001–1010) does not. POST only checks: can_access(model, 'add', create_kwargs_check) can_access(model, 'copy_related', obj) For JobTemplate, can_add (awx/awx/main/access.py:1465–1520) gates on inventory.use_role + project.use_role + execution_environment.read_role — resource-level roles that do not imply read on the source JT — and can_copy_related (1522–1534) checks only credentials.use_role. None of these imply the caller can read the source JT.4d CVE-2026-760897.7 HIG

14.0%

——4Formie is a Craft CMS plugin for creating forms. Prior to 2.2.23 and 3.1.31, Formie's formie/sent-notifications/get-resend-modal-content control panel action in SentNotificationsController::actionGetResendModalContent accepts a request-supplied notification ID without permission or object-level authorization checks. Any authenticated user able to invoke the action can enumerate notification IDs and read recipient headers and complete HTML email bodies containing submitted form data, even without the sent-notification viewing permission. This issue is fixed in versions 2.2.23 and 3.1.31.27d CVE-2026-760878.2 HIG

22.8%

——7Formie is a Craft CMS plugin for creating forms. Prior to 2.2.23 and 3.1.31, Formie's anonymous formie/submissions/submit action in SubmissionsController::actionSubmit trusts a client-supplied submissionId when loading an incomplete submission without session binding, ownership validation, or a valid submissionEditToken. An unauthenticated attacker can enumerate sequential IDs and overwrite or hijack another user's in-progress multi-page or save-for-later submission, and the modified data can be persisted and forwarded through notifications or integrations when the submission is completed. This is an incomplete remediation of CVE-2026-47266 because that earlier change validated edit tokens for save-submission but did not protect submit. Completed submissions are excluded by the isIncomplete filter. This issue is fixed in versions 2.2.23 and 3.1.31.26d CVE-2026-760868.5 HIG

20.4%

——6Formie is a Craft CMS plugin for creating forms. Prior to 2.2.23 and 3.1.31, Formie's formie/integrations/form-settings control panel action in IntegrationsController::actionFormSettings is reachable without the required form integration permissions and passes request-supplied settings to a configured integration. An authenticated attacker can replace outbound host properties such as apiUrl while the server uses stored API keys or OAuth tokens, causing non-blind server-side requests to an attacker-controlled or internal host and returning the remote response. This residual flaw remained because the permission gate added in version 3.1.28 excluded the form-settings action. Sites that permit low-privileged or front-end user authentication can therefore expose integration credentials and internal network responses. This issue is fixed in versions 2.2.23 and 3.1.31.22d CVE-2026-751317.8 HIG

7.0%

——2NetworkManager-l2tp through 1.52.4, fixed in 1.52.6, contains a privilege escalation vulnerability that allows local users with permission to create VPN connections to execute arbitrary code as root by injecting pppd options through a crafted VPN username. Attackers can embed a double-quote character or whitespace in the username to break out of the pppd options file quoting context and include the pppd plugin directive, causing the privileged pppd process to load an attacker-controlled shared object.26d CVE-2026-714653.1 LOW

10.6%

——3RunAdHocCommand.build_args() appends limit as bare positional (args.append(limit)) instead of using args.extend(['-l', limit]) like RunJob. A limit beginning with - is parsed as an ansible CLI option. Currently limited to short-circuit flags (--version, --help) since injected element displaces required pattern positional. Would escalate if ansible-core ever defaults pattern.26d CVE-2026-714643.1 LOW

10.6%

——3LaunchConfigurationBaseSerializer.scm_branch has no validate_scm_branch() leading-dash check, unlike Project/JobTemplate/JobLaunch serializers. Schedule and WFJT Node accept --upload-pack=/bin/id as scm_branch. Currently blocked at runtime by jobs.py:1502 ValueError check (defense-in-depth), but the API validation gap means sole reliance on a task-layer guard. Refactoring that guard away would promote this to RCE.26d CVE-2026-714632.7 LOW

18.5%

——6Notification template Jinja AST whitelist only inspects static Getattr nodes. Dynamic subscripts (job['job'+'_env']) and {% if job.id > 100 %} conditional gating bypass both the AST check and the test-render (stub has small job.id). At runtime, the gated branch executes and exceptions write full tracebacks into notification body, which is POSTed to attacker-controlled webhook URL. Leaks install paths, Python version, source line numbers.25d CVE-2026-714624.1 MED

16.0%

——5StringListPathField.to_internal_value() calls os.path.exists() on unbounded user-supplied paths. 200 vs 400 response reveals existence of arbitrary absolute paths on the controller-web pod. Tenant superuser can confirm /etc/tower/SECRET_KEY, k8s service-account token, receptor sockets, ConfigMap mount points. Mainly impactful on managed AAP (ansiblecloud.com) where tenant admin != host admin.27d CVE-2026-714614.3 MED

10.6%

——3HostList.list() catches bare Exception and returns str(e) verbatim. Via host_filter, any authenticated user triggers Django FieldError (leaking complete Host model relation graph including internal reverse accessors) or PostgreSQL DataError (leaking raw database error strings). Two primitives: credential__search=x dumps ORM schema, name__regex=[bad reflects PostgreSQL errors.4d CVE-2026-714604.3 MED

11.5%

——3/api/v2/config/ is protected only by IsAuthenticated. license_info (account_number, subscription_id, pool_id, sku, support_level, instance counts) returned to any authenticated user. The superuser/auditor gate only covers project_base_dir/project_local_paths/custom_virtualenvs, not license_info. Enables social engineering against Red Hat support and estate sizing reconnaissance.27d CVE-2026-714595.0 MED

19.0%

——6JobJobEventsChildrenSummary view has no model/parent_model. ModelAccessPermission.check_get_permissions() falls through (returns True) for any authenticated user. The view uses raw get_object_or_404(Job, pk) without DRF object-level permission check. Zero-privilege user reads event tree structure, event_processing_finished status, and enumerates Job IDs platform-wide via 200/404 oracle. Sibling endpoint /jobs/{id}/job_events/ correctly returns 403.25d CVE-2026-714585.0 MED

21.6%

——6URLModificationMiddleware resolves named-URL lookups against unfiltered Model.objects before RBAC. The 403→404 shim only rewrites 403 responses, leaving the pk=0 miss path with a different 404 detail string. Differential "Not found." vs "No <Model> matches..." reveals whether a named resource (org, credential, inventory, host) exists anywhere on the platform. Enables cross-tenant internal hostname enumeration.27d CVE-2026-63132—

41.3%

——12OpenBao is an open source identity-based secrets management system. Prior to 2.6.0, OpenBao's handleLogicalRecovery path in http/logical.go compared the highly privileged recovery token with ordinary string equality. A remote unauthenticated attacker able to make repeated recovery mode requests and measure response timing could infer the recovery token. The recovered token could then authorize recovery mode operations that read or modify OpenBao data. This issue is fixed in version 2.6.0.21d CVE-2026-63131—

27.7%

——8OpenBao is an open source identity-based secrets management system. Prior to 2.6.0, OpenBao's vault/policy/acl.go could evaluate a broader wildcard ACL grant before more-specific trailing-wildcard ACL paths with capabilities = ["deny"] for a LIST operation. When a parent path permitted LIST and a child path was denied, the trailing-slash lookup could therefore allow listing the denied path. Other operation types are outside the repository advisory's affected scope. This issue is fixed in version 2.6.0.21d CVE-2026-618147.5 HIG

45.9%

——14Jawn is an open source JSON parser. Prior to 1.7.0, Jawn's AsyncParser can perform quadratic work when a single JSON token is delivered across many small chunks because each absorb call rescans the incomplete token from the start. A remote attacker who controls untrusted JSON input and its chunk sizes can exhaust CPU resources and cause denial of service in applications using AsyncParser. This issue is fixed in version 1.7.0.27d CVE-2026-616957.5 HIG

46.7%

——14Wire provides gRPC and protocol buffers for Android, Kotlin, Swift, and Java. Prior to 6.4.1 and 7.0.0-alpha04, Wire's Swift runtime ProtoReader.skipGroup(expectedEndTag:unknownFieldsWriter:) accepts a negative length for a LENGTH_DELIMITED field inside an unknown START_GROUP field. ProtoReader.readData() forwards the negative count to ReadBuffer.readData(count:), whose upper-bound-only check permits the value to reach Foundation Data(bytes:count:) and trigger an unrecoverable process trap instead of a catchable ProtoDecoder.Error. Any Swift process decoding untrusted protobuf bytes can be crashed without authentication, user interaction, or knowledge of the target schema. This issue is fixed in versions 6.4.1 and 7.0.0-alpha04.22d CVE-2026-599907.5 HIG

48.9%

——15Jawn is an open source JSON parser. Prior to 1.7.0, Jawn parse methods accept arbitrarily deep JSON array and object nesting without a depth limit, allowing a remote attacker who can submit untrusted JSON to grow parser contexts until the JVM heap is exhausted. The resulting java.lang.OutOfMemoryError is a fatal Scala error that is not ordinarily handled by scala.util.Try or cats.effect.IO, causing denial of service. This issue is fixed in version 1.7.0.27d CVE-2026-556324.3 MED

20.8%

——6GoCD is a continuous deliver server. From 20.2.0 until 26.1.0, the internal pipeline structure API used for autocompletion while editing pipeline, template, environment, and user-preference configuration returns its users-and-roles mode to regular authenticated users without requiring an administrator role. A lower-privileged user can enumerate configured user names and available role names, which can facilitate attacks against those users. The response does not reveal which roles are assigned to each user, and the endpoint cannot modify data. This issue is fixed in version 26.1.0.27d CVE-2026-55456—

—

——0Rejected reason: This CVE is a duplicate of another CVE.27d CVE-2026-52744—

35.6%

——11GoCD is a continuous deliver server. From 20.2.0 until 26.1.0, the internal GoCD UI fetch-artifact auto-suggestion API at /go/api/internal/pipelines/**/upstream does not adequately authorize access to upstream dependency data. An authenticated user can retrieve inter-pipeline dependency hierarchy details and user-defined pipeline, stage, job, and artifact plugin reference names that the user cannot otherwise view in the UI. The endpoint is read-only and does not permit modification of pipeline data. This issue is fixed in version 26.1.0.22d CVE-2026-91775—

30.5%

——9LimeSurvey Community Edition 7.0.14 fails to safely encode attacker-controlled content from a crafted .lss survey file when displaying import warnings, resulting in XSS in the administrative interface.18d CVE-2026-888405.3 MED

9.9%

——3BusyBox TLS get_client_hello() reads past the end of the input buffer when parsing a truncated ClientHello message.24d CVE-2026-888396.7 MED

1.9%

——1BusyBox passwd/group tokenize() references a stale endpoint pointer after trimming, causing an out-of-bounds write of heap pointers.25d CVE-2026-888376.5 MED

18.4%

——6BusyBox httpd treats yescrypt ($y$) password hashes as plaintext during Basic Authentication, inverting the authentication check.25d CVE-2026-888356.1 MED

1.6%

——0BusyBox dpkg read_package_field() steps past a NUL terminator on malformed .deb packages, causing an out-of-bounds heap read.25d CVE-2026-888315.3 MED

13.7%

——4BusyBox httpd IP deny rules with invalid CIDR prefix lengths fail open, leaving a parsed IP with a zeroed mask so the rule matches no clients.25d CVE-2026-869387.3 HIG

0.8%

——0A DLL hijacking vulnerability in the FileMaker Pro installer for Windows allowed a local user to execute arbitrary code with elevated administrator privileges by placing a malicious DLL file in the installer directory. This vulnerability is addressed in FileMaker Pro version 26.0.3.15d